Field of the Invention
The invention relates to a notification system. More particularly, the invention relates to systems and methods of managed biometric-based notification.
Discussion of the Related Art
Retail crime, shoplifting, and fraud are increasing in the United States. This increase brings a rising incidence of violence. In 2006, it was reported that the value of property lost in such cases, excluding shoplifting, was $18 billion. Shoplifting resulted in a $40 billion loss that same year, totaling $58 billion in 2006. Some suggest that part of the problem has been the “professionalism” of retail theft, which connotes well organized individuals or gangs stealing large quantities of merchandise. Many surveillance systems are currently being used to combat such crime, but the most effective systems are those which utilize biometric components.
Biometrics is the study of methods for characterizing and recognizing an individual based upon physical and behavioral traits, or in other words, a system which uses measurable biological properties to identify individuals. Physical traits are those related to the characteristics of the human body, e.g. fingerprints, iris geometry, and face recognition. Behavioral traits are those linked to a signature, voice or keystroke. It is because biometrics measures qualities that an individual cannot change, that it is most effective for authentication and identification purposes.
Biometrics are an increasingly important part of the tool set used to securing a wide range of retail stores, facilities, areas, information, and environments. Use of biometric-based identification systems increases because such systems can provide substantially more security than many traditional security systems (e.g., usernames, passwords or personal identification numbers). Face recognition technologies can be used, for example, to determine whether an individual is permitted entry into a home, office, or similar environment, or to determine if an individual is wanted.
Generally, in what is known as “one-to-many match,” and prior to a biometric feature extraction process, an analog or digital representation of biometric characteristics are obtained from a biometric capture device. Many conventional systems rely on digital imaging technologies to capture data, which can include optics, a camera, or other electronic equipment. The digital representation of the images (i.e., the raw or unprocessed image data) is then processed by an algorithm that converts the image data into a particular representation (i.e., a biometric marker or template).
Biometric features are information processed or extracted from a biometric sample or samples, which can then be used for comparison with a stored biometric reference. From the recognition sample, the biometric feature extraction creates a template which is compared with one or multiple biometric templates from a database. Due to the statistical nature of biometric samples, there is generally no exact match possible. For that reason, the decision process will only assign the biometric data subject to a biometric template and confirm recognition if the comparison score exceeds an adjustable threshold. Face recognition works by using a computer to analyze an individual's facial structure. The biometric software takes a number of points and measurements, including the distances between characteristics such as eyes, nose and mouth. This may also include angles of certain features such as the jaw and forehead, and lengths of various portions of the face.
Research today focuses on developing software or algorithms for biometric identification. The developing algorithms aim to reduce a known problem in the art, namely, the high numbers of false positives and negatives, which are called the False Acceptance Rate (FAR) and the False Recognition Rate (FRR), considered Type I and Type II errors in statistical models. The technologies involved in biometric identification include Segmentation, Decomposition Methods, namely, Eigenface, Local Feature Analysis (LFA), and Independent Component Analysis (ICA), and also include Support Vector Machines, Elastic Bunch Grapes, Implicit 3-D models and methods.
While more accurate and precise algorithms are essential to the identification of an individual, the implementation of current systems generally result in an inability to properly disseminate information. Furthermore, current systems are deficient in their implementation in a real-time environment using current technology network components. This deficiency stems from a lack of control over the environment, user integration and management over the system network.